Friday, May 8, 2009

Attacking "Crusaders" Against Conflicts of Interest with Logical Fallacies

An editorial appeared in the latest edition of Nature Medicine that suggested "crusaders" are making much too much fuss over insignificant "competing financial interests." [Anonymous. A really serious conflict. Nature Medicine 2009; 15: 463-4]

The Editorial's Arguments

Instead of regarding all conflicts of interest as part of a "monolithic scourge," it suggested we should only attack them in two specific situations.

Bad Conflicts of Interest

The first is when it leads to falsification of research data:

If there is one reason we need to worry about CFIs [competing financial interests], it is because they could lead to misconduct whenever a researcher has a financial incentive to fabricate data. So, if a clinical trial goes ahead on the basis of fraudulent preclinical work or if a drug is approved on the basis of false clinical data, the problem is more profound than can be addressed by the declaration of a CFI. Researchers or clinicians who produce sham data to continue receiving funds from a company must be subject to severe disciplinary actions from their institutions and funding bodies.

The second is when it leads to withholding of data in certain circumstances:

Companies that conceal negative data from regulators to receive approval for their drug should face not only fines but also penalties as serious as criminal charges. In any case, the focus should move away from CFI declarations and toward exposing and punishing scientific misconduct.

Some Conflicts of Interest are Unavoidable

On the other hand, the editorial asserted that conflicts of interest are unavoidable in certain situations.

But even though a lot has been written about this subject, it is seldom acknowledged that not all CFIs are equally insidious and that some are even necessary for the success of translational research.

Scientists who share their expertise with a company and clinicians who agree to conduct a clinical trial have to be compensated for their work in the same way that every other professional ought to be rewarded for his or her labor. To suggest that they should make their contributions freely is disingenuous, and to argue that they should not get involved at all can only slow the development of new medicines.

Academic institutions and their employees must be free to benefit financially from the fruits of the advances made in their laboratories, either by independently developing them into products or by selling them to the highest bidder. In the US, the Bayh-Dole Act has protected this right for nearly three decades, and CFIs that result from a university's right to enact this policy should not raise eyebrows.

Bad Outcomes of the "Crusade" Against Conflicts of Interest

In addition, the editorial argued that because of excess attention paid to minor or unavoidable conflicts of interest, bad things may happen. Clinical trials would be discouraged.
However, the fact remains that clinical trials are very expensive, and only the pharmaceutical sector has both the incentive and the deep pockets to pay for them.

And innovation would be slowed.

A third issue that is rarely acknowledged is that the stridency with which the CFI crusaders fight financial interests has created a milieu in which pharmaceutical firms have increasingly less breathing space to conduct their business, leading to a further slowing of drug development.

Finally, the editorial exhorted:


Let us instead accept that most financial interests do not represent a conflict as a matter of course and that the influence of money is negative only if it leads to scientific fraud—the one infidel we need to burn at the crusader's stake.
Logical Fallacies and Inflammatory Rhetoric

It is discouraging to see such sloppy thinking and inflammatory rhetoric appear under the masthead of a prominent journal. How was the thinking sloppy? Let me count the ways, and explain the logical fallacies used by the editorial to support its contention.

Arguing that Conflicts are Unavoidable Using False Dilemmas and Appeals to Common Practice

The editorial suggests some conflicts of interest may be unavoidable, even necessary for the advancement of science. To do this, it provided a series of false dilemmas and appeals to common practice. A false dilemma is an argument that x and y are mutually exclusive, when they are not. An appeal to common practice is an argument that because something is common practice, it is necessary, moral, correct, or the only course possible.

The editorial argued that researchers would not work for free, implying that if drug and device companies could not pay researchers to perform clinical trials, no clinical trials would be done. This is a false dilemma, because it is based on the unstated assumption that there is no other possible source of payment for clinical trial researchers other than the companies whose products the trials evaluate. It may be true that it would be hard to get anyone to conduct clinical trials for free, but there is no law of nature that dictates that all clinical trials must be financed directly by companies whose products are the subjects of these trials, or that prevents clinical trials from being funded by government agencies and/or non-governmental organizations disinterested in whether the results favor one treatment or another. Such disinterested sponsors paying researchers would not create conflicts of interest.

The editorial argued that if academic institutions and their employees cannot benefit financially from "the fruits of the advances made in their laboratories," there would be no drug and device development. This is a false dilemma, because it is based on the unstated assumption that the only setting for drug and device development is academic institutions. There is no natural law, however, that restricts drug and device development to academic institutions, or prevents drug development to be done by drug and device companies, or by independent laboratories. The editorial also appended an appeal to common practice. Bayh-Dole may be the current law in the US, but it is hardly natural law.

The editorial argued that without the "deep pockets" of the "pharmaceutical sector" and device companies, no one would pay for clinical research. This was a repetition of the false dilemma noted above. There is no natural law that prevents government and non-governmental organizations from sponsoring clinical trials. The editorial again appended an appeal to common practice. Current laws and social conventions have made drug and device companies extremely profitable, but continuation of their profitability is hardly inevitable.

Red Herrings to Distract from Other Bad Effects of Conflicts

While the editorial identified two particularly bad kinds of conflicts of interest, it ignored other instances that could be equally bad. This was something of a red herring. Although the two examples of bad effects of conflicts of interest were undeniably important, they distracted from other bad effects which the editorial chose to ignore.

The editorial correctly identified falsification of data as bad, but ignored the entire catalog of ways in which clinical research studies can be manipulated to make them more likely to favor a sponsor's product. See our posts (here and here) and the links contained therein for multiple examples of such manipulation.

The editorial noted that concealing data from regulators in order to win drug approval is bad, but ignored all the instances in which data has been suppressed outside of the regulatory process to make drugs or devices seem more efficacious or less dangerous than they really are. See this link for some of our posts containing relevant examples. If it is bad to fool people who must make regulatory decisions, why is it not so bad to fool physicians and patients who must decide whether to use drugs or devices?

The editorial further ignored that conflicts that affect how research is conducted or reported may be linked to conflicts that affect how research is disseminated or used in decision making. The clinical researcher who is paid by a drug or device company to conduct a clinical trial may also be an academic physician paid by the same company as a consultant or through a speakers' bureau to give talks and presentations, and write review articles favorable to the company's product. If it would be bad for the researcher to falsify data to make the product look better, why would it be acceptable for the researcher to act as a huckster for the product while pretending to be a disinterested academic? See this link for posts discussing stealth marketing, and this one for posts discussing the use of key opinion leaders.

Ambiguity and Ad Hominem Attacks Inflating the Bad Effects of Fighting Conflicts

While the editorial minimized the bad outcomes of conflicts of interest, it resorted to ambiguity and inflammatory language to exaggerate what might happen were conflicts to be reduced.

Its most dire prediction of the results of reducing conflicts was that doing so would create a "milieu" in which "pharmaceutical firms have increasingly less breathing space." Literally, this statement was nonsense. Metaphorically, it did not make a clear argument that these highly profitable companies (and even more profitable device companies) could not survive in a "milieu" in which their products were tested by researchers which the companies could not influence. This was not clearly a logical fallacy, but just word-play presented a reasoned argument.

Given how prevalent political correctness is, repeatedly characterizing people concerned about conflicts of interest as "crusaders" is inflammatory indeed. Being called "pharmascolds" is one thing, but "crusader" in some circles is regarded as equivalent to racist, a nasty term to label those who promote the separation of clinical research and medical education from commercial interests. Thus, the editorial winds up with a rather nasty version of an ad hominem attack.

We have seen numerous examples of the use of logical fallacies in defense of conflicts of interest, often by authors who were themselves conflicted. It is especially sad to see the editors of a scholarly medical journal employ such logical fallacies in the defense of conflicts of interest.

ADDENDUM (11 May, 2009) - See also the comments by Dr Doug Bremner in the Before You Take That Pill Blog.

ADDENDUM (20 May, 2009) - For an example of how the term "crusader" has become politically incorrect in the US context, see this article in the Boston Globe.